Synaptic target selection in Drosophila Abstract: Genetic screens in Drosophila identified many of the cell-surface and secreted (CSS) proteins that are intensively studied today as regulators of axon guidance in both vertebrate and invertebrate systems. This proposal describes a genetic screen for CSS proteins that function as synaptic target labels in the embryonic/larval neuromuscular system. This system is ideal for examination of target labeling mechanisms, because it contains only 36 motor neurons and 30 muscle targets and has an invariant innervation pattern. Each identified motor neuron innervates a specific muscle fiber. Although many genes that regulate axon guidance in this system have been identified, we know very little about how individual muscle fibers are recognized as targets by motor axons. To address this problem, we first defined CSS proteins that cause axonal mistargeting when they are overexpressed on all muscle fibers. We did this by constructing a database of all genes in Drosophila that encode CSS proteins likely to be involved in cell recognition events. We then searched through all the existing collections of UAS (GAL4 binding site)-containing ('EP-like') element lines to find insertions immediately upstream of these CSS genes that could be used to confer tissue-specific, high-level expression by crossing them to GAL4 "driver" lines. We obtained EP-like insertions that can drive 410 of the 979 genes in the database, or over 40% of the putative cell recognition repertoire. We crossed each line to a pan-muscle GAL4 driver and examined F1 progeny larvae by antibody staining and confocal microscopy. We found 30 genes whose expression on all muscles causes high-penetrance axonal mistargeting phenotypes but does not perturb muscle structure. Six of the genes are in a specific family encoding proteins with extracellular domains containing leucine-rich repeats (LRRs), which are protein interaction modules. This proposal describes experiments to assess the functions of four LRR proteins that are expressed in muscles and appear to function as synaptic target labels, and to determine if the LRR family encodes additional target labels. The first specific aim concerns the Tartan (Trn) and Capricious (Caps) proteins. Loss-of- function phenotypes for trn and caps suggest that they function in a partially redundant manner in the embryo. In larvae, selective expression of Trn or Caps on muscle 12 only produces alterations in targeting specificity. We will determine the loss-of-function (LOF) larval phenotypes generated by knockdown of both Trn and Caps in a single muscle or in all muscles. We will also attempt to develop a method for labeling single motor axons in larvae, so that we can observe how genetic perturbations affect targeting of individual identified axons. Specific aims 2 and 3 concern two "new genes", CG14351/haf and CG8561. We have used genetic and RNAi analysis to show that the proteins encoded by these genes are required for the normal innervation of ventrolateral muscles. We will make null mutations in these genes and conduct a genetic interaction screen to find components of the CG14351/Haf signaling pathway. We will also determine whether CG8561, the ortholog of a mammalian IGF-1 binding protein, is a component of the insulin/IGF-1 signaling pathway. The final specific aim describes experiments to examine the entire LRR family to determine if it encodes other muscle target labels. To do this, we will make UAS-cDNA constructs and obtain or make RNAi lines for 41 LRR genes and assess their phenotypes in larvae. For all genes producing phenotypes, we will then make a map of their expression patterns in muscle fibers during the period of axonal outgrowth. This information will allow us to begin to combine LRR protein perturbations, knocking down multiple genes on specific muscles, in order to examine whether muscle fibers are labeled for targeting by expression of specific ensembles of LRR proteins.